Installation¶

Poetry¶

poetry add opsml

Pip¶

pip install opsml

Optional Dependencies¶

OpsML is designed to work with a variety of 3rd-party integrations depending on your use-case.

Types of extras that can be installed:

Postgres: Installs postgres pyscopg2 dependency to be used with OpsML
```
poetry add "opsml[postgres]"
```
Server: Installs necessary packages for setting up a Fastapi based OpsML server
```
poetry add "opsml[server]"
```
GCP with mysql: Installs mysql and gcsfs to be used with OpsML
```
poetry add "opsml[gcp_mysql]" or "opsml[gcs,mysql]"
```
GCP with postgres: Installs postgres and gcsgs to be used with OpsML
```
poetry add "opsml[gcp_postgres]" or "opsml[gcs,postgres]"
```
AWS with mysql: Installs postgres and s3fs dependencies to be used with OpsML
```
poetry add "opsml[aws_mysql]" or "opsml[s3,mysql]"
```
AWS with postgres: Installs postgres and s3fs dependencies to be used with OpsML
```
poetry add "opsml[aws_postgres]" or "opsml[s3,postgres]"
```
Azure with mysql: Installs postgres and s3fs dependencies to be used with OpsML
```
poetry add "opsml[azure_mysql]" or "opsml[azure,mysql]"
```
Azure with postgres: Installs postgres and s3fs dependencies to be used with OpsML
```
poetry add "opsml[azure_postgres]" or "opsml[azure,postgres]"
```

Example setup for gcs storage and postgres with opsml server¶

  poetry add "opsml[gcp_postgres, server]"

Environment Variables¶

OpsML requires 1 or 2 environment variables depending on if you are using it as an all-in-one interface (no proxy) or you are using it as an interface to interact with an OpsML server.

OPSML_TRACKING_URI: This is the sql tracking uri to your card registry database. If interacting with an OpsML server, this will be the http address of the server. If this variable is not set, it will default to a local SQLite connection.
OPSML_STORAGE_URI: This is the storage uri to use for storing ml artifacts (models, data, figures, etc.). OpsML currently supports local file system, Google Cloud Storage, AWS S3 and Azure. If running OpsML as an all-in-one interface, this variable is required and will default to a local folder if not specified. If interacting with an OpsML server, this variable does not need to be set.

Storage Type	URI Format
Local	`./path/to/folder`
Google Cloud	`gs://bucket_name`
AWS S3	`s3://bucket_name`
Azure	`az://container_name`

Example setups¶

Server setup with local storage and sqlite:

export OPSML_TRACKING_URI="sqlite:///opsml.db"
export OPSML_STORAGE_URI="./opsml_registries"

Server setup with gcs storage and postgres (with ssl):

export OPSML_TRACKING_URI="postgresql+psycopg2://username:password@opsml-host:5432/db_name?connect_timeout=5&sslmode=require"
export OPSML_STORAGE_URI="gs://opsml_bucket"

Server setup with s3 storage and mysql:

export OPSML_TRACKING_URI="mysql+pymysql://username:password@opsml-host:3306/db_name"
export OPSML_STORAGE_URI="s3://opsml_bucket"

Server setup with azure storage and mysql:

export OPSML_TRACKING_URI="mysql+pymysql://username:password@opsml-host:3306/db_name"
export OPSML_STORAGE_URI="az://opsml_container"

TLDR Scenarios¶

Server is already setup and I need to interact with it from the client side (notebook, python script, cli, etc.):

Set OPSML_TRACKING_URI to the http address of the server

I need to setup the Server:

Set OPSML_TRACKING_URI to the sql tracking uri of your card registry database
Set OPSML_STORAGE_URI to the storage uri of your choice
Follow additional instructions in server docs